Rapid translation of clinical guidelines into executable knowledge: a case study of COVID‐19 and on‐line demonstration

Fox, J; Khan, O; Curtis, H.; Wright, A; Pal, C; Cockburn, N; Cooper, J; Chandan, J.S.; Nirantharakumar, K

The Polyphony programme is a rapidly established collaboration whose aim is to build and maintain a collection of current healthcare knowledge about detection, diagnosis and treatment of COVID‐19 infections, and use Artificial Intelligence (knowledge engineering) techniques to apply the results in patient care. The initial goal is to assess whether the platform is adequate for rapidly building executable models of clinical expertise, while the longer term goal is to use the resulting COVID‐19 knowledge model as a reference and resource for medical training, research and, with partners, develop products and services for better patient care. In this Polyphony progress‐report we describe the first prototype of a care pathway and decision support system that is accessible on OpenClinical.net, a knowledge sharing repository. Pathfinder 1 demonstrates services including situation assessment and inference, decision making, outcome prediction and workflow management. Pathfinder 1 represents encouraging evidence that it is possible to rapidly develop and deploy practical clinical services for patient care and we hope to validate an advanced version in a collaborative internet trial. Finally, we discuss wider implications of the Polyphony framework for developing rapid learning systems in healthcare, and how we may prepare for using AI in future public health emergencies. This article is protected by copyright. All rights reserved.